Magic: The Gathering is a tabletop and digital collectible card game created by Richard Garfield released in 1993 by Wizards of the Coast. Currently is one of the most famous trading card game in the world.

Modern is one of the most popular format in the competitive setting of Magic: The Gathering.

It allows all cards printed from July 29, 2003 to the present to be played (with some exceptions).

A player must use a deck consisting of at least 60 cards plus 15 cards in the sideboard and it can use at most 4 copies of the same card, except for Basic Land cards that can be played in any number.

The analysis was carried out on all decks that ranked well in competitive tournaments with at least 8 players in the month of April 2022

The dataset that I have analyzed is one made by myself. I have downloaded 798 deck lists from mtgtop8 in txt format and I have arranged them in order to obtain a dataset where each entry has the following variable values:

• The deck_name
• The player
• The deck_list
• The type

The variable deck_list is a nested dataframe where an entry contains the information about 1159 different cards (got by scryfall):

• the name of the cards played in the deck
• the number of copies played
• the colors
• and other data

• Find out the best players in the month of April and see if they are specialized in specific strategies or if they are versatile.

• Find out the most played cards and see how they relate with each other and with the deck played, and then the same for the less used cards and for signature cards.

From an analysis of the played decks, it can be seen that the most played deck type is the AGGRO.

Despite the most popular deck type, the most popular deck is 4c Control of the the type CONTROL which is the less popular deck type.

Anyway more than the 40% of the 10 most played decks is an AGGRO deck.

The best players are the players that figure out more than once in the dataset. This is due to the fact that they have obtained a good result in more than a tournament in April.

Apparently it was not easy to get more than a good result in April, as a matter of fact only 59 players out of 729 got more than one good result. And of these 59 only 10 got three.

In this section I have considered the 10 players with three results.

What I wanted to find out is whether to be a good player it was possible to play different types of decks or whether it was better to specialize on one strategy.

Some interesting results:

• Richard_Neal has played two deck types.
• Bosko_Milakovic is the only one who managed to use all the three deck types. So I will say he’s the most versatile player.
• All the other players are specialized on a single deck type.

So the tendency is to be specialized in a single deck type.

Speaking about the single decks, I have analyzed all the player with more than one result.

The relation must be interpreted as: x has played the same deck as y.

• There are only two players who have played three different decks: Bosko_Milakovic and Ryan_Gassaway
• Bosko_Milakovic is not only the most versatile player but he’s also the most creative one. In fact he has played three decks that none of the other best players have played.
• Jesse_James_Adams is the player with the highest number of links as he has played the two most popular decks: 4c Control and UR Aggro.

In term of different decks (that may be different but belong to the same type), the players are a bit less specialized with a good number of them which has played different deck.

This graph shows the 20 most played non land cards. Quite surprisingly there are no black cards between them. However, it was predictable because the most played deck (4c Control) uses all colors except black.

To find the first black card we must look at the 25th position: Thoughtseize , anyway it is the only one black card in the top 40.

How often the top cards are used together?

It results that almost all the top cards are used together at least in one the deck. The only exception is Shardless Agent which has non links with 9 of the other top cards.

Apparently the use of the top cards is bind to the popularity of the deck.

There are 142 cards that figure out in a single copy in all the analyzed decks.

• There are 46 communities of single-copy-cards.
• 12 of them are composed of a single elements: they are the single-copy-cards which are played with no others single-copy-cards.
• Enchantress is the deck that uses the greatest number of single-copy-cards: 10.

The tendency to play single-copy cards is more typical of less popular decks while top decks tend to maintain more standard lists.

The fact that a very popular deck plays single-copy-cards, in the case of 4c Control which plays 5, is given by the fact that the deck is widely played and therefore it is easier to find different lists.

The signature cards are cards that are played in a single deck but in any amount of copies. There are 579 signature cards but not all of them are interesting: in fact cards that are played in a deck which appear only once in the dataset are not so important.

For this analysis I have considered only the most important signature cards: cards that are played in decks that appear at least twice and are played in average at least 3 copies per deck. For a total of 83 signature cards.

The results shows that the use of signature cards in not driven by the popularity of decks but by their type, in fact it seems that it is a feature more proper of the COMBO decks.